DESCRIPTION (Applicant's abstract): The annual cost of depression in the United States has been estimated to be greater than 40 billion dollars, and includes the direct costs of treatment and indirect costs related to lost productivity. The elderly are particularly prone to increased disability from depression. Prospective studies demonstrate that up to two-thirds of elderly depressed patients have a poor outcome in one year, and over one-fourth are re-hospitalized in the year following the acute depressive episode. Yet the current public policy debates have focused on the cost of entitlement programs such as Medicare and have ignored the complex relationship between the cost of the initial treatment, efficacy and long term benefit in reducing overall healthcare costs. The primary aim of the present proposal is to develop practical guidelines for the acute and prophylactic therapy of treatment resistant geriatric major depression. In this proposal, severely depressed elderly patients will be evaluated, treated with either medication or electroconvulsive therapy (ECT) and followed for 18 months. The data generated will be used to develop statistical models of the most clinically effective and cost-efficient acute, continuation and maintenance treatments for severe geriatric depression,. 360 patients meeting DSM-IV criteria for Major Depression, severe (approximately half with psychotic features) and the American Psychiatric Association guidelines for ECT will be randomized to either an acute trial of ECT or a 6 week prospective medication trial. Medicare costs, neuropsychological, sociodemographic and clinical data will be analyzed to determine both the costs of an additional medication trial, and the patient characteristics which are associated with a response to medication/ECT. Patients who fail the prospective medication trial will be administered an acute course of ECT. The second and third phases of the study will evaluate alternative prophylactic therapies after an acute response to ECT. Patient who are severely depressed and respond to ECT. Patients who are severely depressed and respond to an acute course of ECT have extremely high healthcare costs and attendant morbidity and mortality. The risk of relapse in 6 months in these patients is unacceptably high (50-70%) using the conventional strategy of placing the patient on continuation antidepressant medication. The 180 patients who respond to an acute course of ECT or fail the prospective medication trial and then respond to an acute course of ECT and will be randomized to either 6 months of continuation strategies. In the third phase, 160 patients will be followed for an additional year in order to examine the long term benefits of alternative continuation strategies (i.e., ECT vs. medication) on measures of mood stability, neuropsychological function and total healthcare costs. The data from these studies will be used to develop a clinical checklist which can be used in general psychiatric practice to assist clinicians in making clinically effective and cost-efficient treatment decisions.